Issue with metacyc and minpath

[ato...@gmail.com]

Hello,

I recently started using Humann2 and I seem to be having some issues downloading the full program with all dependencies (even with the pip install).

As such, this command

$ humann2 --input $SAMPLE --output $OUTPUT_DIR

returns the error:

ERROR: You are using the demo ChocoPhlAn database with a non-demo input file. If you have not already done so, please run humann2_databases to download the full ChocoPhlAn database. If you have downloaded the full database, use the option --nucleotide-database to provide the location. You can also run humann2_config to update the default database location. For additional information, please see the HUMAnN2 User Manual.

After several other similar errors, I was able to run the following command successfully, just linking to each db:

humann2 --input ./file.fastq.gz --nucleotide-database /location/chocophlan/ --protein /location/uniref/ --metaphlan /location/metaphlan2/ --output /output/

However my final output was not three files, but only one "file_genefamilies.tsv"
not pathcoverage or pathabundance.

I think this is likely because the program is not detecting MinPath and Metacyc, however I noticed on the forum that these files are not available to download. I'm not sure how to locate these files from the initial pip install. This could be an issue with my machine, but is there any way to get these files for me to link locally on my computer?

Thank you,
Andrew

[Lauren McIver]

Hi Andrew - Are you seeing errors in the HUMAnN2 log that it is unable to find MinPath? If you could attach the log or just part of it, that would be great. Also are you running in conda or another environment where it might be setup to install from a wheel instead of from source? If so, this could explain why you are not seeing the dependencies being installed. Add the options listed in the "Getting Started with HUMAnN2" section ( http://huttenhower.sph.harvard.edu/humann2#markdown-header-getting-started-with-humann2 ) to install from source.

Thanks,

Lauren

[ato...@gmail.com]

Thank you! it was an install issue after all!

I have a follow up question: in section 3.3 of the tutorial it says
"Using the humann2_rename_table script, attach names to the unnamed level-4 EC features."

I used the script

$ humann2_rename_table -i merged_families.level4ec-cpm.tsv -o merged_fa
milies.level4ec-cpm-names.tsv -n uniref90_level4e

but got the kickback:

usage: humann2_rename_table [-h] [-i INPUT]
[-n {infogo1000,metacyc-rxn,uniref90,kegg-module,ec,go,metacyc-pwy,pfam,eggnog,uniref50,kegg-pathway,kegg-orthology}]
[-c CUSTOM] [-s] [-o OUTPUT]
humann2_rename_table: error: argument -n/--names: invalid choice: 'uniref90_level4ec' (choose from 'infogo1000', 'metacyc-rxn', 'uniref90', 'kegg-module', 'ec', 'go', 'metacyc-pwy', 'pfam', 'eggnog', 'uniref50', 'kegg-pathway', 'kegg-orthology')


I didn't get this issue when using the regroup table with ref90_level4ec. I recently updated my databases, so I don't think that is it.

Any ideas?

[Lauren McIver]

Hello Andrew, I am glad to hear the reinstall solved your issue. Some of the utility scripts have database files that are large enough that we can not include them in the humann2 pip package. Follow step #6 in the installation instructions to install these utility databases: http://huttenhower.sph.harvard.edu/humann2#markdown-header-getting-started-with-humann2 . Once these are installed you should see more options for the utility scripts names.

Thanks,

Lauren

[ato...@gmail.com]

Thank you for the prompt reply Lauren!

I downloaded the databases earlier in the week but just did so again to be sure. It's still not working properly. Am I entering the -n correctly?

$ humann2_rename_table -i merged_families.level4ec-cpm.tsv -o

merged_families.level4ec-cpm-names.tsv -n uniref90_level4ec

I'm still getting this back

usage: humann2_rename_table [-h] [-i INPUT]
[-n {infogo1000,metacyc-rxn,uniref90,kegg-module,ec,go,metacyc-pwy,pfam,eggnog,uniref50,kegg-pathway,kegg-orthology}]
[-c CUSTOM] [-s] [-o OUTPUT]
humann2_rename_table: error: argument -n/--names: invalid choice: 'uniref90_level4ec' (choose from 'infogo1000', 'metacyc-rxn', 'uniref90', 'kegg-module', 'ec', 'go', 'metacyc-pwy', 'pfam', 'eggnog', 'uniref50', 'kegg-pathway', 'kegg-orthology')

Thank you,
Andrew

[Lauren McIver]

Hi Andrew - The name option for ecs for that specific utility script is just "ec". Did you want to add names or regroup abundances? If you want to add names to a file of ec abundances use humann2_rename_table with "-n ec". If you have a table of uniref90 abundances that you would like regrouped to ec abundances use the humann2_regroup_table utility with "-g uniref90_level4ec". Sorry it is a little confusing that the two options differ a bit. The reason they differ is because there are multiple options of abundances to ecs for the regroup script.

Thanks,

Lauren

[ato...@gmail.com]

Thank you! That was it! I was looking to add names to ec abundances, so -n ec was exactly what I was looking for!

Andrew